A lane departure warning system is used to give a warning to drivers who are in a sleepy, fatigue-driving or attention-deficit state, so as to prevent the occurrence of traffic accidents due to lane departure.
Currently, the lane departure warning systems adopted in foreign countries mainly include an AURORA system, an AutoVue system, a Mobileye_AWS system and a Driver Support System (DSS), and the lane departure warning systems adopted in China mainly include a JLUVA-1 system, and an embedded, Digital Signal Processor (DSP)-based lane departure warning system. A conventional lane departure warning model is mainly based on one of the followings: a position of a vehicle in a current lane, a coming departure amount of the vehicle, and a duration within which the vehicle is running on a boundary of the lane. These methods are mainly used to accurately determine a distance between a lane marking and the vehicle by measuring the position of the vehicle in the lane. At this time, it is necessary to define a camera, so as to establish a geometrical imaging model consisting of the camera, the vehicle and a road surface. However, the establishment of the geometrical imaging model may be readily affected by such factors as a type of the camera, a type of a lens, an installation position of the camera, a type of the vehicle and a type of a road.
Hence, there is an urgent need to provide a scheme so as to overcome the above-mentioned defect.